压缩感知视角的鲁棒波达角估计算法  

Compressive sensing perspective for robust DOA finding

作　　者：王友华[1,2] 张建秋[1] 

机构地区：[1]复旦大学电子工程系,上海200433 [2]模拟集成电路重点实验室,重庆400060

出　　处：《系统工程与电子技术》2014年第1期1-8,共8页Systems Engineering and Electronics

基　　金：国家自然科学基金(61171127);模拟集成电路重点实验室基金(9140C090110130C09003)资助课题

摘　　要：从压缩感知的视角对鲁棒波达角估计进行了探索,通过将可能存在的波达角进行空间离散化,从而将波达角估计问题转换为压缩感知信号支撑恢复问题。同时将阵元存在的增益失配、相位失配和阵元间互耦等非理想因素,通过一阶近似,将其建模成均值为理想流形矩阵的随机矩阵,从而建模了阵列非理想特性和波达角空间离散化带来的误差。基于这种新的随机测量矩阵模型,提出了一种基于压缩感知的鲁棒波达角估计算法,分析表明本文提出算法对阵列模型扰动和角度空间离散化具有良好的鲁棒性。仿真验证了分析结果。A robust direction of arrival （DOA） estimation method under the framework of compressive sensing is proposed in this paper. The proposed algorithm firstly discretizes the candidate angle space. Then the problem of DOA estimation is converted to the sparse signal recovery by utilizing the property that DOA corresponds to the support of sparse signal. The array steering matrix is modeled by a random error matrix with its mean as the idea1 steering matrix through first order approximation when considering non-idealities of sensor array and basis mismatch during angle discretization. Based on the new random steering matrix, a new compressive sensing based DOA estimation algorithm is proposed. Theoretical analysis proves the robustness of the proposed method. Numerical simulations conducted validate the analytical results.

关 键 词：阵列信号处理 鲁棒波达角估计 模型误差 信号恢复 压缩感知 

分 类 号：TN95[电子电信—信号与信息处理]